You are helping me build the following AI agent workflow.

## Goal
Hacker News Trends to Telegram: Every four hours, fetches Hacker News via Algolia, scores trends by velocity, and sends a clean HTML digest to Telegram.

## Specification
- What it does: You receive a clean HTML digest of trending Hacker News stories on your Telegram every four hours. The system automatically scores trends by velocity, filters out low-engagement posts, and optionally translates summaries before delivering the results to you.
- Trigger: On a schedule (Scheduled · every 4 hours)
- Autonomy: Fully hands-off
- Expected setup effort: about an afternoon
- Tools/services involved:
  - HTTP Request: Fetch data from Algolia API
  - Code: Calculate Gravity Score and filter noise
  - AI Agent: Translate titles and summaries
  - OpenRouter Chat Model: Provide LLM capabilities for translation
  - Telegram: Send formatted digest to chat

## Known pitfalls, handle each one explicitly in your implementation
1. Telegram message length exceeds 4096 characters easily with multiple posts, so split long digests into separate messages.
2. OpenRouter model latency may delay the digest delivery, consider caching summaries or using a faster model endpoint.
3. HTML tags in post titles can break the Telegram formatting, strip all HTML entities before sending to the API.

## Reference implementation
https://n8n.io/workflows/12748 (workflow template)
Fetch and inspect this before building. If it matches my stack, adapt it;
if not, rebuild the pattern with my tools.

## Process requirements
1. Before building: ask me which of the listed tools I actually use and
   what my platform is (n8n / Make / code / other). Do not assume.
2. Adapt the pattern to my answers; do not force the reference stack.
3. Address every pitfall above; tell me how you handled each.
4. Provide a test plan I can run before letting this touch real data.
5. Ask before any step that sends messages, modifies data, or spends money.

Source: https://usecasesforagents.com/use-case/hacker-news-trends-to-telegram/ via usecasesforagents.com